Abstract: An apparatus and a method for the disclosure relates to an artificial intelligence (AI) system that simulates functions such as recognition and determination of the human brain by using a machine training algorithm such as deep learning and an application of the AI system are provided. A neural network training method includes obtaining target modality signals of a first domain aligned in a time order and auxiliary modality signals of a second domain that are not aligned in the time order, extracting characteristic information of the target modality signals using a first neural network model, estimating the time order of the auxiliary modality signals using a second neural network model, and training the first neural network model based on a result of the estimation and the characteristic information.

Abstract: Provided are an artificial intelligence (AI) system using a machine learning algorithm such as deep learning, and applications thereof. A method, performed by an electronic apparatus, of processing images includes obtaining a plurality of training image sets corresponding to a plurality of types of target objects, wherein training images in the training image sets are labeled with feature points forming a preset structure, generating a first artificial intelligence (AI) model for determining a standard structure based on the labeled feature points, by using the training images in the training image sets, identifying a face in a training image transformed based on the standard structure, and training a second AI model for verifying the first AI model, based on an image regressed from the transformed training image, and the training image before being transformed.

Abstract: An apparatus and a method for the disclosure relates to an artificial intelligence (AI) system that simulates functions such as recognition and determination of the human brain by using a machine training algorithm such as deep learning and an application of the AI system are provided. A neural network training method includes obtaining target modality signals of a first domain aligned in a time order and auxiliary modality signals of a second domain that are not aligned in the time order, extracting characteristic information of the target modality signals using a first neural network model, estimating the time order of the auxiliary modality signals using a second neural network model, and training the first neural network model based on a result of the estimation and the characteristic information.

Abstract: Disclosed herein is a method of processing images based on image segmentation using higher-order correlation clustering. In an image segmentation method according to an embodiment of the present invention, an input image is segmented into superpixels. A hypergraph is constructed by connecting two or more adjacent superpixels, among the superpixels, to one another. A joint feature map is created by extracting feature vectors from respective edges of the hypergraph, and partitioning the hypergraph based on higher-order correlation clustering in consideration of specific constraints.

Abstract: A direction of arrival (DOA) estimation device and method are provided, in which the DOA estimation device includes a sensor unit configured to detect a signal and comprising two or more sensors to output sensor signals as a detect signal in response to the detected signal, and a controller configured to calculate statistical distribution data indicative of statistical distribution of each of the sensor signals outputted from the two or more sensors, respectively, retrieve statistical distribution data indicative of statistical distribution of source signal which is non-stationary signal entrained in the signal of the calculated statistical distribution data, and estimate DOA of the source signal based on the retrieved statistical distribution data.

Abstract: Disclosed are a method and system for producing a multimedia fingerprint based on quantum hashing. The method includes receiving an input of a multimedia file, extracting a quantum hash type fingerprint from the input multimedia file, calculating similarity between the extracted quantum hash type fingerprint and a binary fingerprint stored in a database, and selecting, as a calculation result, data having a fingerprint calculated as having the highest similarity.

Abstract: A continuous phonetic recognition method using semi-Markov model, a system for processing the method, and a recording medium for storing the method. In and embodiment of the phonetic recognition method of recognizing phones using a speech recognition system, a phonetic data recognition device receives speech, and a phonetic data processing device recognizes phones from the received speech using a semi-Markov model.

Abstract: Disclosed herein is a method of processing images based on image segmentation using higher-order correlation clustering. In an image segmentation method according to an embodiment of the present invention, an input image is segmented into superpixels. A hypergraph is constructed by connecting two or more adjacent superpixels, among the superpixels, to one another. A joint feature map is created by extracting feature vectors from respective edges of the hypergraph, and partitioning the hypergraph based on higher-order correlation clustering in consideration of specific constraints.

Abstract: Enhanced method for embedding watermarks based on integer-to-integer wavelet transform is provided. The method according to the present invention includes the steps of: (A) dividing an original image (X×Y) to a plurality of image blocks (M×N); (B) selecting image blocks for embedding an location information that indicates image blocks to be watermarked; (C) embedding the location information into the image blocks selected in the step (B); and (D) embedding watermarks into remaining image blocks which are not selected in the step (B).

Abstract: Disclosed are a method and system for producing a multimedia fingerprint based on quantum hashing. The method includes receiving an input of a multimedia file, extracting a quantum hash type fingerprint from the input multimedia file, calculating similarity between the extracted quantum hash type fingerprint and a binary fingerprint stored in a database, and selecting, as a calculation result, data having a fingerprint calculated as having the highest similarity.

Abstract: Enhanced method for embedding watermarks based on integer-to-integer wavelet transform is provided. The method according to the present invention includes the steps of: (A) dividing an original image (X×Y) to a plurality of image blocks (M×N); (B) selecting image blocks for embedding an location information that indicates image blocks to be watermarked;(C) embedding the location information into the image blocks selected in the step (B); and (D) embedding watermarks into remaining image blocks which are not selected in the step (B).

Abstract: Provided are an apparatus for and a method of determining a transmission rate in speech transcoding. An input frame is classified as speech or silence based on a first threshold value that is predetermined for at least one of a fixed code-book gain value, an adaptive code-book gain value, a noise to signal rate, and a pitch delay that correspond to an input parameter of a coded bit stream. An input frame classified as voiced is classified as stationary or non-stationary based on a third threshold value that is predetermined for the amount of change in the ACBG value or a difference between the minimum and maximum pitch delays. An input frame, classified as voiced by a voiced/unvoiced classifying portion, is classified as voiced or non-stationary based on a class of a previous frame.